CGAL users discussion list

Hello,
After some challenge with MS .NET 2003, I decided to compile in linux. the 
specifications are:

CGAL 3.2
OS:          i686_Linux-2.6    
Compiler:    GNU 3.4.2                                
Support for: BOOST, X11, GMP, MPFR, CORE, ZLIB, and QT.


by using "make" I get some errors in a file, one of them is:
../../include/CGAL/Polygon_Voronoi_diagram_2.h:331: error: there are no 
arguments to `finite_edges_begin' that depend on a template parameter, so a 
declaration of `finite_edges_begin' must be available
../../include/CGAL/Polygon_Voronoi_diagram_2.h:331: error: (if you use 
`-fpermissive', G++ will accept your code, but allowing the use of an 
undeclared name is deprecated)
 the line it says is:

typedef Segment_Delaunay_graph_hierarchy_2
      <Sdg_traits_2, CGAL::Tag_true, Sdg_ds_2>           Sdg_2;
...
void get_edges (std::list<Edge_sites_2>& edges) const 
{ 
    edges.clear(); 
     // Go over all finite arcs in the dual graph.  
    typename Sdg_2::Finite_edges_iterator eit; 
    for (eit = finite_edges_begin()/*EERROORR*/; eit != 
finite_edges_end()/*EERROORR*/; ++eit) 
        _compute_primal_edge (*eit, edges); 
}

An intresting thing is the same finite_edges_begin() is used in "Alpha" 
example but with different definition and without compilation error. I mean 
in "Alpha_shape_3.h" it is:
  
typedef typename Dt::Finite_cells_iterator    Finite_cells_iterator;
....
Finite_edges_iterator eit = finite_edges_begin();
for( ; eit != finite_edges_end() ; ++eit){
if (classify(*eit, alpha) == type) *it++ = *eit;
...

what is wrong with finite_edges_begin() or GCC? if installation menu shows 
that it supports for GCC 3.4, so why it does have problem?


Thanks,